Study Of Reactive Power Optimization Of Distribution Network Considering Short-Term Load Forecasting

The reactive power centralized dispersive optimization control idea based on short-term load forecasting in power distribution networks is proposed in this dissertation. The short-term load forecasting in each substation based on artificial neural networks is carried out firstly, then the forecasted results are transferred to the automatic voltage control (AVC) system in a local power system. AVC can solve the global reactive power optimization based on genetic algorithms, obtain the optimized results for each substation, and unify the coordination control for each substation's operation. This scheme combined the advantages of the global optimization, the centralized control and the distributed control. This overcame the limitation for local reactive power / voltage optimization control in substations, increased voltage qualification rate, and reduced the power loss. This method has been verified in Jinan distribution network and achieved good practical effect.